The overall aim of this study is to provide a comprehensive picture of the course and outcome of four specific personality disorders (PDs): schizotypal (STPD), borderline (BPD), avoidant (AVPD), and obsessive-compulsive (OCPD). The present proposal continues a multi-site collaborative effort to follow a carefully diagnosed sample of 668 subjects having either these representative PDs or major depressive disorder (MDD) (controls) for the period from 2 to a minimum of 6 years after recruitment. Using a prospective, longitudinal, repeated measures design, we will develop the same basic knowledge about course and outcome for the PDs that has previously resulted from similar investigations of affective and anxiety disorders, thus addressing an important gap in our knowledge. The extended period of follow-up is essential to discern clinically meaningful descriptions of course and outcome and their determinants. The sample is large enough and sufficiently diverse demographically to attain a unique array of results generalizable to most clinical settings. To accomplish our overall aim, we propose three approaches: I. descriptive, II. predictive, and III. validating. The descriptive approach will provide data on diagnostic stability of PDs, on presence and course of comorbid Axis I disorders, on persistence of functional impairment, and on utilization of health care resources that allow comparison between the PDs and to similar data on Axis I disorders. The predictive approach will identify clinically meaningful determinants of prognosis within and across PDs. The validating approach will examine the homogeneity of descriptive and longitudinal features for the PDs, as defined by the categorical DSM system, and how this compares with alternative dimensional schemes.